Note that Mir's input/gesture targeting code is designed to avoid these problems, by sending input events only to the surface where the gesture began. But I'm guessing mir_demo_server doesn't use that existing window stack logic.
Note that Mir's input/gesture targeting code is designed to avoid these problems, by sending input events only to the surface where the gesture began. But I'm guessing mir_demo_server doesn't use that existing window stack logic.